Estimating Child Mortality: An Alternative Modeling Approach

Leontine Alkema, National University of Singapore
Wei Ling Ann, National University of Singapore

Millennium Development Goal 4 calls to reduce the under-five mortality rate by two thirds between 1990 and 2015. The United Nations Inter-Agency Group for Child Mortality Estimation (IGME) estimates child mortality in every country to measure progress. This is a challenging task for many countries because of issues with data quality and quantity. For the majority of countries, IGME uses a multiple spline regression estimation approach, which can give inaccuracies in the estimated rate of decline over time. We propose to use an alternative estimation approach, in which a time series model describes the rate of decline. Furthermore, we propose to estimate the model parameters for each country with a Bayesian hierarchical model, such that trends within countries are estimated based on the country's own history as well as observed trends in all other countries.